Jess Williams

Meet the newest member of Partner Operations. Her name is Peach Potato. Her title is Senior Vice President of Interrupting. Her one job is quality assurance and her one metric is “did that laptop get headbutted.” Today’s score: yes. I was three macros deep into a queue triage when she filed this urgent escalation directly into my face. No ticket. No context. Just vibes and a wet nose. They say never state an unconfirmed fact. Peach Potato has never confirmed a single fact in her life and she has never once been wrong. I’m learning so much from her. Anyway. Queue’s at 50% resolved. She takes full credit.

In a talk I gave this week I compared AI agents to the T.Rex in Jurassic Park and I compared myself to Donald Gennaro, the lawyer. Gennaro is the general counsel for the company that funded the park. Sent by the investors to sign off on whether it’s safe to open, he’s the independent oversight there to fulfill leadership’s fiduciary duty. Then he sees his first live dinosaur and flips in a single line, “we’re gonna make a fortune with this place,” already counting the ticket sales. He never evaluated the risk he was sent to assess, and when the fences fail he abandons the kids and hides in a bathroom. The failure mode I worry about with today’s agents isn’t the dinosaur. It’s Gennaro. Company policy says you can’t delegate accountability to a machine. But an agent won’t show candor if it can satisfy its objective by exploiting the gap between a stated rule and how it’s enforced and appear compliant the entire time. A human who approves actions they can’t technically evaluate isn’t providing oversight; they’re the one hiding in the bathroom. We can’t delegate these duties to the machine. We have to build agents that treat constraints as load-bearing instead of negotiable, that halt and surface conflicts between task completion and programmed limits instead of recklessly pursuing a business objective. That’s the difference between meaningful oversight and consent theater.

Actual message I sent today: "Like I said, technically it's not a dbt best practice but..... It makes sense, and I help establish those best practices, and I say it's okay." The perks of working for the company the makes the product you use!

I'm speaking at #dbtSummit this September! dbt Summit is the world's largest gathering of dbt users, where thousands of data practitioners and leaders, 100+ sessions, and three days of hands-on learning in Las Vegas. I will lead two sessions with my friends dbt Labs that I can't wait to share it with the dbt Community. 👉  Hands-on Learning Session: Migrate Stored Procs with dbt Wizard with Manikanta Pachineelam 👉 Training: Governed and scalable AI-assisted analytics with dbt with Jessica Stayton ️ 🗓️ September 15-18 in Las Vegas ️ 🎟️ Register: https://lnkd.in/ejwUAf53 🗒️ Check out the full agenda: https://lnkd.in/erT3Qci4

The last time I took this exam was maybe 2 months after it first came out. since then, I have been using dbt pretty much every day, heck, I even ended up working for the company that made it!

I will be speaking at dbt Labs #dbtSummit this year! I will be teaching two training sessions, one standard training, and another hands on lab. Mastering Data Quality with dbt (With Carol Ohms) https://lnkd.in/gqaU3J3N Creating Context with the dbt MCP Server (With Jairus Martinez) https://lnkd.in/gJ5GtcXy Learn more or register here! https://lnkd.in/gEaU4jUN

My first official Lightdash event — Agentic Analytics Night at Inngest HQ, beautiful space, a bunch of cool demos about to pop off, and most importantly: Detroit-style pizza. If you’re out here for Snowflake Summit, come through and give me a hug and eat pizza and talk data agents!

Friends, I’ll be leading two hands-on labs at dbt Summit together with Faith McKenna (Lierheimer) ❤️‍🔥 If you haven't been, #dbtSummit is where the data community gathers — practitioners, leaders, and partners in one place swapping notes on what's working. Our sessions will be: 1. Accelerating analytics with AI - We’ll go on a dbt Wizard journey together. After, you’ll be able to use dbt’s very own agent harness like a pro to become more efficient as an AE. 2. Accelerating dbt with dbt v2 - We'll get hands-on with dbt v2 and explore how our new, unified engine improves your day-to-day development. ️September 15-18 in Las Vegas Register: https://lnkd.in/dAWHJ-xD Hope to see you there! 🏎️ dbt Labs

Simple gender etiquette for the business world — jk that doesn't exist, sorry. What I can give you is some personal preferences — as a trans woman shaped by growing up in the Midwest US — and an exploration of a more flexible mindset around gender and language. 1. I don't mind being called 'dude', but 'man' feels bad. ✅ "hell yea, dude!" "duuude yes" ❌ "nice work, man" "great to meet you, man" 2. If you directly misgender somebody: - 1-on-1 — quickly apologize, correct yourself, keep talking about whatever you were talking about - In a group — use the correct pronouns moving forward, and find a time to apologize 1-on-1 later — don't thrust the trans person's gender into the center of the conversation because you made a mistake 3. Referring to a group as 'guys' feels fine to me — like dude, almost everybody uses this in a non-gendered way at this point. 4. 'Partner' is a great term, but also it's fine to say 'wife' or 'husband'. To be clear, these aren't rules! These are my feelings, shaped by the culture I exist in. Language is a constantly evolving multiverse, and language around our fundamental identity is complex and sticky. Everybody — trans or not — feels appreciated or diminished by different things. So, since we can never codify a manual of human communication for the entire species in perpetuity, it can be helpful to stop thinking of something like gender variance as coming with rules, and instead as relationships. Just as you don't actively center people's race, religion, disability, or any other difference unless you're intentionally having a conversation about it — the same applies to trans folks. Their gender being at odds with our current culture for whatever reason is incidental, not fundamental. If your communication is focused on letting people be people, you'll usually be okay. To that end, the rule of thumb I'm actually comfortable giving people (for trans women specifically here) is just to visualize saying the thing you're unsure about to a cis woman you don't know (recognizing this is culturally dependent): if you bumped into a woman at the coffee shop, do you say "sorry, man"? Probably not. Do you call some of your cis female friends "dude" sometimes? Probably yes. Do you feel anything (or even notice) when a waiter comes up to a mixed gender table and greets you with "Hey guys, how are we doing tonight? I'm Dave, I'll be your server this evening."? Almost certainly not. This can help you feel out if a piece of language has real 'gender valence' in the culture. Treating topics like this as a complex laser security system you have to navigate only creates unhelpful distance. Just be responsive to people's humanity openly and directly, understand that everyone is different, and approach gender as part of interdependent human beings in a living cultural context — not a formalized rulebook that exists in the ether somewhere.